(closed) student assistant

On Reinforcement Learning, Generative Models, and Music


The Machine Learning Research Lab (Munich) is looking for a student assistant (f/m/d) to work on generative models and reinforcement learning applied to music. The internship is a continuation of previous work on improving the expressiveness of Variational Auto-Encoders (VAE) and the structure of their latent spaces. These models were already successfully applied to interactive musical improvisation and were deployed in a live session to play with a human drummer.

The focus of this internship is to keep improving the quality of the musical generative model and to further explore the interplay between Reinforcement Learning (RL) algorithms for musical composition.

Your Tasks

  • train deep VAEs on musical data (sequential data) and investigate the inferred representations
  • develop RL algorithms to generate music
  • benchmark different recent Neural Network architectures such as Transformers and Graph Neural Networks.

Your Qualifications

You should meet a few of these and be eager to learn about the rest.

  • student (bachelor/master) in a STEM-related discipline;
  • interest in machine learning, generative models, sequence models, reinforcement learning;
  • very good with Python; experience with Tensorflow or Pytorch is a plus;
  • very good with numerical optimisation, probability theory, information theory, calculus and linear algebra;
  • good knowledge of tools such as git.

Please contact us at arg-min \(\text{@}\) argmax.ai if you are interested. Mention "RL for music internship" in your email subject for this particular opening.

Announcement at Volkswagen Stellenbörse